A dsRNA Viral Transcriptional Regulator Evades Innate Immunity by Hijacking Host CoTranscription Factor DHX9

Figures

• 

  FIGURE 1. DHX9 is a host factor that interacts with σ3. (A) Purified proteins visualized by Coomassie blue staining. In HEK293T cells, Protein complexes associated with pNTAP‐σ3 were isolated by tandem affinity purification, separated by SDS‐PAGE, and visualized by Coomassie blue staining. Molecular weight marker is indicated on the left. The pNTAP empty vector served as a negative control. * indicates the bait protein σ3; (B) σ3 interacts with DHX9 in an DNA/RNA‐independent manner. HEK293T cells were transfected with either empty vector or pcDNA3.1‐Flag‐σ3. After 48 h, cells were harvested. Cell lysates were treated with or without RNase A (100 ng mL−1), Super Nuclease (25 U mL−1), or RNase III (10 U mL−1) at 37 °C for 20 min and then subjected for co‐IP; (C) The interaction between σ3 and DHX9 increases when cell lysates were treated with increasing amounts of poly I:C. Cell lysates from HEK293T cells were collected at 48 h post‐transfection and were then incubated with either poly I or poly I:C at the indicated concentrations on ice for 30 mins. After incubation, lysates were subjected for co‐IP using anti‐Flag agarose beads. (D) DHX9 interacts with σ3, but not σNS during REOV…

• 

  FIGURE 2. Characterization of the interaction between σ3 and DHX9. (A,B) σ3 mutants K287T, R208T, and K293T did not interact with DHX9. (C) Space‐filling surface view of the σ3 dimer structure (PDB ID: 1FN9) showing the mutated residues. Mutated residues in Red indicates disrupted the σ3‐DHX9 interaction. Residues colored yellow are dsRNA‐binding defective but are capable of binding DHX9. (D) Schematic representations of full‐length and the indicated truncations of DHX9. HD: Helicase domain; RBD: dsRNA binding domain; RGG: C‐terminal RGG‐containing domain. Numbers denote amino acid residues. (E) The helicase domain of DHX9, but not the RBD or RGG domains, interacts with σ3. Myc‐tagged DHX9 (full length or truncated) was transfected together with Flag tagged σ3 into HEK293 cells. Co‐IP was performed 48 h post‐transfection using anti‐Flag beads, followed by immunoblotting to detect the indicated proteins. (F) Full length DHX9 or its helicase domain (HD) interact with the σ3 R296T mutant.

• 

  FIGURE 3. DHX9 acts as a host restriction factor for REOV through its helicase activity. (A,B) T1L‐WT and T1L‐R296T viruses replicated more efficiently in DHX9‐deficient A549 cells, whereas T1L‐K287T did not. A549 cells were first infected with lentivirus encoding shRNA against DHX9. After 48 h, cells then infected with the indicated REOV at MOI 10. Viral growth was shown as changes in viral titer from 0 to 24 h pi. The efficiency of DHX9 knockdown was verified by western blotting in (A). Data shown represent the mean ± s.d. of three independent experiments. (C,D) Viral growth of T1L‐WT was assessed after reconstitution of either WT DHX9 or its helicase dead mutant K417R in A549 cells deficient in DHX9. The expression levels of DHX9 were examined by western blotting in (C). Viral growth was shown as changes in viral titer from 0 to 24 h pi. Data shown represent the mean ± s.d. of three independent experiments. Multiple non‐paired t test was used to analyze differences (ns = no significant, *p< 0.05, **p< 0.01).

• 

  FIGURE 4. σ3 inhibits DHX9 helicase activity in vitro. (A) Coomassie Blue staining of recombinant σ3 (WT/K287T) and DHX9 purified from 293F cells. (B) DHX9 resolves R‐loops in vitro. The in vitro helicase assay on R‐loops was performed by incubating the 5’ Cy5.5‐labeled R‐loop substrates with increasing amounts of recombinant DHX9 at 37 °C for 10 min, and subsequently analyzed by native polyacrylamide gel electrophoresis and fluorescence imaging. (C) Quantification of DHX9's unwinding activity. Error bars indicate SD as derived from three independent experiments. (D) σ3 WT, but not K287T, inhibits the helicase activity of DHX9. DHX9 was incubated with increasing amounts of either σ3 WT or K287T at room temperature for 30 min, and then subjected for helicase assay by adding R‐loop substrates and ATP. (E) DHX9, but not σ3 binds R‐loops. Indicated amounts of DHX9 or σ3 were incubated with R‐loop substrates at 37 °C for 15 min. The reaction was then terminated and loaded onto native PAGE gels to check the band shift. All images were collected using Licor Odessey imaging system.

• 

  FIGURE 5. σ3 inhibits TNF‐α‐stimulated NF‐κB gene expression through its interaction with DHX9. (A) REOV inhibits TNF‐α‐stimulated NF‐κB gene expression. HEK293T cells were infected with T1L at MOI 10. At 18 h pi, cells were treated with or without TNF‐α at 10 ng mL−1 for 1 h. RNA was extracted from cells, and the levels of CXCL2, NFKBIA and UBE2Z mRNA relative to GAPDH were measured by RT‐qPCR and comparative ΔΔCT analysis. (B‐G) The mRNA levels of RELB and IL5 relative to GAPDH were measured by RT‐qPCR after σ3 expression (B and E), or DHX9 knockdown (C and F), or DHX9 overexpression (D and G) in HEK293T. Data are presented as the mean ± s.d. of three independent experiments. Non paired t test was used to analyze the differences (ns = no significant, *p< 0.05, **p< 0.01, ***p<0.001).

• 

  FIGURE 6. Wild‐type σ3, but not K287T, impairs both the DHX9‐dependent recruitment and the pause‐release of RNA polymerase II (Pol II). (A) DHX9 helicase activity is required for the increased interaction between DHX9 and RNA Pol II upon TNF‐α treatment. Co‐IP experiment was performed in HEK293T between exogenously expressed Flag‐tagged DHX9 (either WT or K417R) and endogenous RNA pol II following TNF‐α treatment. The co‐IP efficiency was quantified by comparing the band intensity in the IP lane to the band in the input lane. All samples were normalized to mock (set at 1). Quantification of three independent experiments were shown in (B); (C) expression of σ3 WT, but not K287T, significantly reduced the interaction between DHX9 and RNA pol II upon TNF‐α treatment. The co‐IP efficiency was quantified by comparing the band intensity in the IP lane to the band in the input lane. All samples were normalized to mock (set at 1). Quantification of three independent experiments were shown in (D). (E) Average Pol II CUT & Tag signal profile on genomic loci (defined as 3 kb upstream of annotated TSS to 3 kb downstream of annotated TES) in HEK293T cells expressing σ3 or K287T upon TNF‐α treatment…

• 

  FIGURE 7. REOV infection, or σ3 expression, leads to an upregulation of cellular R‐loops, potentially contributing to viral replication. (A) Venn diagram showing the overlap of σ3 interactome (green) with known interactors of R‐loops (blue). (B) Cellular R‐loop levels were upregulated upon treatment of TNF‐α, IFN‐β, or REOV infection. A549 cells were treated with IFN‐β (10 ng mL−1), TNF‐α (10 ng mL−1) for 3 h or infected with REOV at MOI 50 for 6 h. Genomic DNA were then extracted, and R‐loop levels were detected using slot blot assay with the S9.6 antibody. (C) Quantification of three independent experiments as shown in (B). Relative S9.6 band intensity was first normalized to methylene blue staining, and then R‐loop level in mock was set as 1. (D,E) Knocking down DHX9 led to an upregulation in cellular R‐loop levels in Mock (D) or cells treated with TNF‐α (E). A549 cells were infected with lentivirus encoding shRNA against DHX9. At 48 h pi, cells were either mock‐treated or treated with TNF‐α. Genomic DNA isolated from nuclear extracts were analyzed for R‐loops using slot blot assay with the S9.6 antibody. (F) σ3 expression led to an upregulation in cellular R‐loop levels. Band intensity…

• 

  FIGURE 8. σ3 expression, leads to an upregulation of R‐loops at the promoter proximal region of TNF‐α‐responsive genes. (A) Average R‐loop (upper) and H3K4me3 (lower) CUT & Tag signal profile on genomic loci (defined as 3 kb upstream of annotated TSS to 3 kb downstream of annotated TES) in HEK293T cells expressing σ3 upon TNF‐α treatment. (B) Genome‐wide distribution of R‐loop (upper) and H3K4me3 (lower) peaks. (C and D) Comparison of abundance of R‐loop (C) and H3K4me3 (D) at promoter proximal region of TNF‐α‐responsive genes or TNF‐α‐non‐responsive genes. (E) Heatmaps show the read density that is ranked by decreasing occupancy of R‐loop (upper) and H3K4me3 (lower). (F,G) ChIP‐qPCR analysis of R‐loop (F) or H3K4me3 (G) enrichment at the promoter proximal region of RELB, CXCL3, and ACTB in HEK293T cells expressing σ3 upon TNF‐α treatment. Enrichment is calculated by percent input method. Data are presented as the mean ± s.d. of three independent experiments. Paired t test was used to analyze the differences (ns = no significant, *p< 0.05, **p< 0.01, *** p<0.001). (H) A working model of how σ3 inhibits NF‐κB gene expression (created using BioRender.com): ① σ3 competes with Pol II for…